EViews.com

Dear QMS Gareth,
I am trying to estimate a gravity model using the fixed and random effects model but the issue is that I have an unbalanced trade data set which I will describe as follows:
The importing country is the group ID but I am having problems on how to specify an elaborate cellid to accomodate the periodic effect because in my dataset, some of the importing countries are have multiple observation for some years.

I have gone through the Eviews Manual for Eviews 6 that I am using for the analysis but such a special case was not treated in the Manual.
I request for your assistance as regards arranging my dataset to be able to accomodate both cross section and periodic fixed effect.

Attached to this post is the data structure.

Best Regards

You'll have to create a series that some how uniquely identifies the time period. EViews won't care what you use as long as it uniquely identifies the period.

Dear QMS Gareth,
could you please suggest what to do in this case because I will be having a group to specify the importer, another to specify the period and the third to specify the cellid which uniquely identify the observation.

When using the Menu, How do I specify the identifier series in this case?

Kindly explicitly explain or help fill in the period identifier and the cellid in this case.

Best Regards

As you point out, you can't use date as the period identifier, since there are cases where there are two dates for the same country. Thus you'll have to come up with a different one. Again EViews doesn't mind what you use, as long as it is unique. For example, you could create a column in your Excel file called "Mydate", and put in years like "2009:1", "2009:2" etc...